Vehicle owner indoor positioning method and system based on UWB technology

Abstract

The invention provides a UWB technology-based vehicle owner indoor positioning method and system. The method provided by the invention comprises the following steps: when a positioning demand exists, determining a UWB tag based on a smart phone of a vehicle owner, and sending a positioning command to the UWB tag through the smart phone; after the UWB tag receives the positioning command, pulse signals are sent to a plurality of UWB base stations with known positions, and the distance between the UWB tag and the UWB base stations is determined based on the flight time of the pulse signals; performing iterative optimization on the distance based on a Kalman filtering algorithm to obtain optimized distance information; and the positioning model calculates the position information of the UWB tag based on the optimized distance information and the position coordinates of the UWB base station, and sends the position information to the smart phone of the vehicle owner. The method and the system provided by the invention are used for solving the problems that the existing indoor positioning technology is insufficient in positioning precision in a complex environment, part of methods are complex to operate and the like, and the accuracy and the practicability of indoor positioning are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of indoor positioning, in particular to a car owner indoor positioning method and system based on UWB technology. BACKGROUND

[0002] With the vigorous development of the Internet of Things, people's demand for positioning services is no longer limited to outdoor environments, and the demand for accurate positioning in indoor scenarios is increasingly prominent, such as store navigation in large shopping malls and vehicle searching in underground parking lots, all of which cannot be achieved without reliable indoor positioning support. Therefore, it is of great practical significance to develop indoor positioning.

[0003] At present, indoor positioning technologies mainly include ultra-wideband (UWB), Bluetooth, WiFi, etc. Among them, the positioning accuracy of Bluetooth and WiFi positioning technologies is relatively limited in complex indoor environments due to the influence of multipath effect, signal attenuation, etc.; although the UWB positioning method based on UWB technology has many advantages through the transmission of nanosecond-level extremely narrow pulses, it still has some deficiencies in practical application. The UWB positioning method based on time difference of arrival (TDoA) needs to synchronize the clocks between base stations, which is relatively complex to operate; the single-sided two-way ranging (SS-TWR) method based on flight time (ToF) has a large distance measurement error due to the increase of local clock offset error with the increase of counting time; and the coexistence of line-of-sight (LoS) and non-line-of-sight (NLoS) environments in indoor environments further affects the accuracy of positioning.

[0004] Therefore, there is an urgent need for a method to solve the problems of insufficient positioning accuracy of existing indoor positioning technologies in complex environments, complex operation of some methods, and to improve the accuracy and practicality of indoor positioning. SUMMARY

[0005] Therefore, the present application provides a car owner indoor positioning method and system based on UWB technology to solve the problems of insufficient positioning accuracy of existing indoor positioning technologies in complex environments, complex operation of some methods, and to improve the accuracy and practicality of indoor positioning.

[0006] Specifically, the present application is realized by the following technical solutions:

[0007] The first aspect of the present application provides a car owner indoor positioning method based on UWB technology, which comprises:

[0008] When there is a positioning demand, determine a UWB tag based on the car owner's smartphone, and send a positioning command to the UWB tag through the smartphone; each car owner's smartphone corresponds to one UWB tag, and the UWB tag is the target to be positioned;

[0009] After the UWB tag receives the positioning command, pulse signals are sent to a plurality of UWB base stations with known positions, distances from the UWB tag to the UWB base stations are determined based on the flight time of the pulse signals; the number of the UWB base stations is not less than 3;

[0010] The distances are iteratively optimized based on a Kalman filtering algorithm to obtain optimized distance information;

[0011] A positioning model calculates the position information of the UWB tag based on the optimized distance information and the position coordinates of the UWB base stations, and sends the position information to the smart phone of the car owner.

[0012] The second aspect of the application provides a car owner indoor positioning system based on UWB technology, which comprises a positioning server, a gateway, a plurality of UWB base stations and a UWB tag installed in a smart phone of a car owner;

[0013] The smart phone of the car owner is configured to determine a UWB tag when there is a positioning demand and send a positioning command to the UWB tag; each smart phone of the car owner corresponds to one UWB tag, and the UWB tag is a target to be positioned;

[0014] The UWB tag is configured to respond to the positioning demand of the car owner, send pulse signals to a plurality of UWB base stations with known positions after receiving the positioning command, and determine distances from the UWB tag to the UWB base stations based on the flight time of the pulse signals; the number of the UWB base stations is not less than 3;

[0015] The positioning server is configured to iteratively optimize the distances based on a Kalman filtering algorithm to obtain optimized distance information;

[0016] The positioning server is further configured to calculate the position information of the UWB tag based on the optimized distance information and the position coordinates of the UWB base stations;

[0017] The gateway is configured to receive the distance information fed back by the UWB base stations and forward the distance information to the positioning server, and send the position information calculated by the positioning server to the smart phone of the car owner.

[0018] The UWB technology-based indoor positioning method and system for a car owner provided in the application significantly improve the accuracy, reliability and practicability of UWB tag positioning through full-process collaborative design. First, the one-to-one correspondence between the smart phone and the UWB tag and the sending of positioning commands realize the accurate binding of the positioning target, avoid confusion in the multi-tag scene, ensure the uniqueness of the positioning object, and clarify the target for subsequent processing. Second, the UWB tag sends pulse signals to no less than three UWB base stations with known positions and obtains original distance data by using the high-precision time measurement capability of UWB technology based on time-of-flight ranging. The redundant information provided by multiple base stations meets the basic conditions of triangular positioning, providing a physical basis for positioning calculation. Third, the Kalman filter algorithm iteratively optimizes the distance, smooths the ranging noise through the prediction and update mechanism in time sequence, reduces random errors, and makes the optimized distance closer to the true value, providing more reliable input for the positioning model. Finally, the positioning model calculates the position information based on the optimized distance and the base station coordinates and feeds back to the mobile phone, uses the optimized data to improve the position calculation accuracy, and ensures the real-time accessibility of the results through the mobile terminal output. These means form a closed loop from target association, original data acquisition, noise suppression to final calculation, and together realize the high-precision, anti-interference and practicability of UWB tag positioning in indoor environment, enabling the car owner to accurately and timely obtain the position information of himself or the associated target. BRIEF DESCRIPTION OF DRAWINGS

[0019] Figure 1 A flowchart of the UWB technology-based indoor positioning method for a car owner provided in Embodiment One of the application;

[0020] Figure 2 A schematic diagram of the communication between the UWB tag and the UWB base station provided in the application;

[0021] Figure 3 A structural schematic diagram of the UWB technology-based indoor positioning system for a car owner provided in Embodiment Two of the application. DETAILED DESCRIPTION

[0022] The exemplary embodiments will be described in detail herein with reference to the drawings. When the following description refers to the drawings, the same numbers in different drawings represent the same or similar elements unless otherwise indicated. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the application.

[0023] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting. As used in this application, the singular forms "a," "an," and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "and / or," as used herein, refers to and encompasses any or all possible combinations of one or more of the associated listed items.

[0024] It is to be understood that, although the terms first, second, third, etc. can be used herein to describe various information, these terms are not intended to denote a particular order or hierarchy. These terms are used merely for the purpose of distinguishing between two or more information. For example, without departing from the scope of the present application, a first information can be termed a second information, and similarly, a second information can be termed a first information. The word "if" as used herein means "when" or "upon" or "in response to the determination" depending on the context.

[0025] The specific embodiments are given as follows to introduce the technical scheme of the present application in detail.

[0026] Figure 1 The flow chart of the indoor positioning method of the car owner based on the UWB technology provided in Embodiment One of the present application is shown in FIG. 1. Please refer to Figure 1 The method provided in the present embodiment can include:

[0027] S101, when there is a positioning requirement, determining a UWB tag based on the smart phone of the car owner, and sending a positioning command to the UWB tag through the smart phone.

[0028] The smart phone of each car owner corresponds to one UWB tag, and the UWB tag is the target to be positioned.

[0029] Specifically, the UWB tag is a small device integrating the ultra-wideband (UWB) communication function, which is installed in the smart phone of the car owner, can receive the positioning command sent by the smart phone, and can interact with the UWB base station through pulse signals. The core function of the UWB tag is to serve as a carrier of the target to be positioned, and to provide a basis for positioning through ranging data.

[0030] Further, the smart phone of each car owner is bound to one UWB tag, that is, one car owner corresponds to one UWB tag. The car owner sends a positioning command to the bound UWB tag through the smart phone, triggers the tag to start the positioning process, and the positioning result of the UWB tag will finally be mapped to the position of the car owner, that is, the position of the UWB tag is equivalent to the position of the car owner.

[0031] In a specific implementation, the owner's smartphone is uniquely identified by a device ID, a user account, or a Bluetooth MAC address. Based on a pre-configured mapping relationship (e.g., a database record), the smartphone ID is mapped to the corresponding UWB tag ID. Further, an instruction package containing a positioning request is generated, including a timestamp, request type, etc. The UWB communication module (e.g., a built-in chip or an external accessory) of the smartphone sends the positioning instruction to the mapped UWB tag.

[0032] S102, after the UWB tag receives the positioning command, sends a pulse signal to a plurality of UWB base stations with known positions, and determines the distance from the UWB tag to the UWB base stations based on the time of flight of the pulse signal.

[0033] The number of UWB base stations is not less than 3.

[0034] Specifically, the UWB base station is a device with ultra-wideband signal transceiver function deployed at a fixed position indoors, and its position coordinates are pre-determined and known. It is a signal receiving and processing node in the indoor positioning system, used for communication with the UWB tag and auxiliary calculation of the UWB tag's position. Since the calculation of the UWB tag's position needs to take the position of the UWB base station as the reference, the known position coordinates of the UWB base station as the positioning reference point, combined with the distance information from the UWB tag to each UWB base station, are the core basis for realizing the positioning of the tag.

[0035] Further, the pulse signal is a nanosecond-level extremely narrow electromagnetic pulse sent by the UWB tag to the UWB base station after receiving the positioning command. It is a carrier for transmitting information and measuring distance in UWB technology, and the distance between the UWB tag and the UWB base station is calculated by the propagation time of the pulse signal in space.

[0036] It should be noted that in a three-dimensional space, at least three independent distance parameters (i.e., three-dimensional coordinates x, y, z) are needed to determine the position of a point. Each UWB base station can provide one distance information from the UWB tag to the UWB base station, so at least three UWB base stations are needed to provide three independent distance parameters.

[0037] In the specific implementation, after receiving the positioning command, the UWB tag sends a pulse signal to a plurality of UWB base stations with known positions, and determines the distance from the UWB tag to the UWB base stations based on the time of flight of the pulse signal, including: the UWB tag puts the first timestamp information and tag information when the positioning command is received into a request frame, and sends the request frame to a plurality of UWB base stations with known positions; after receiving the request frame, the UWB base station records the second timestamp information when the request frame is received, puts the third timestamp information after a first time delay and base station information into a response frame, and sends the response frame back to the UWB tag; after receiving the response frame, the UWB tag records the fourth timestamp information when the response frame is received, puts the fifth timestamp information after a second time delay and tag information into a termination frame, and sends the termination frame to the UWB base station; after receiving the termination frame, the UWB base station records the sixth timestamp information when the termination frame is received; each UWB base station calculates the signal propagation time between each UWB base station and the UWB tag according to the respective timestamp information and the tag information, and determines the distance from the UWB tag to the UWB base stations based on the signal propagation time.

[0038] Specifically, the first timestamp information refers to the local clock time information recorded by the UWB tag when receiving the positioning command from the smart phone, and the first timestamp information is put into the request frame together with the tag information of the UWB tag. The second timestamp information refers to the local clock time information recorded by the UWB base station when receiving the request frame sent by the UWB tag. The third timestamp information refers to the local clock time information recorded by the UWB base station after a first time delay (the time from processing the request frame to packaging the response frame) after receiving the request frame, and the third timestamp information is put into the response frame together with the base station information of the UWB base station. The fourth timestamp information refers to the local clock time information recorded by the UWB tag when receiving the response frame sent by the UWB base station. The fifth timestamp information refers to the local clock time information recorded by the UWB tag after a second time delay (the time from processing the response frame to packaging the termination frame) after receiving the response frame, and the fifth timestamp information is put into the termination frame together with the tag information. The sixth timestamp information refers to the local clock time information recorded by the UWB base station when receiving the termination frame sent by the UWB tag.

[0039] In a specific implementation, the UWB communication module of the smart phone receives a positioning command, reads a local clock to generate first timestamp information, encapsulates the first timestamp information and a tag ID into a request frame, and broadcasts the request frame to all UWB base stations with known positions through a UWB wireless channel. The request frame contains the tag ID, the first timestamp information, and a frame type identifier. The UWB base station detects and analyzes the request frame in the UWB signal, records a local clock at the time of receiving as second timestamp information, waits for a time from processing the request frame to encapsulating an answer frame, records a first time delay end time as third timestamp information, encapsulates the second timestamp information, the third timestamp information, and a base station ID into the answer frame, and sends the answer frame to the UWB tag through the UWB channel. The UWB tag detects and analyzes the content of the answer frame, records a local clock at the time of receiving as fourth timestamp information, waits for a time from processing the answer frame to encapsulating a termination frame, records a second time delay end time as fifth timestamp information, encapsulates the fourth timestamp information, the fifth timestamp information, and the tag ID into the termination frame, and sends the termination frame to the corresponding UWB base station through the UWB channel. The UWB base station detects and analyzes the content of the termination frame, and records a local clock at the time of receiving as sixth timestamp information.

[0040] Optionally, each UWB base station calculates a signal propagation time between each UWB base station and the UWB tag according to the respective timestamp information and the tag information, and determines a distance from the UWB tag to the UWB base station based on the signal propagation time, including: calculating a first difference value of the fourth timestamp information and the first timestamp information, the first difference value being a time difference between the UWB tag sending the request frame and receiving the answer frame; calculating a second difference value of the sixth timestamp information and the third timestamp information, the second difference value being a time difference between the UWB base station sending the answer frame and receiving the termination frame; determining a first time delay based on a third difference value of the third timestamp information and the second timestamp information; determining a second time delay based on a fourth difference value of the fifth timestamp information and the fourth timestamp information; calculating a first product of the first difference value and the second difference value, calculating a second product of the first time delay and the second time delay, and calculating a fifth difference value of the first product and the second product; calculating a sum value of the first difference value, the second difference value, the first time delay, and the second time delay, and determining a quotient value of the fifth difference value and the sum value as the signal propagation time; the signal propagation time being a time for a pulse signal to reach the UWB tag from the UWB base station; and determining the distance from the UWB tag to the UWB base station based on a product of the signal propagation time and a propagation speed of electromagnetic waves in air.

[0041] Specifically, Figure 2 A schematic diagram of UWB tag and UWB base station communication is provided in the present application. Please refer to Figure 2, the first difference value is a difference value between the fourth time stamp information and the first time stamp information, representing a total time difference between the UWB tag from sending the request frame to receiving the response frame. The second difference value is a difference value between the sixth time stamp information and the third time stamp information, representing a total time difference between the UWB base station from sending the response frame to receiving the termination frame. The third difference value is a difference value between the third time stamp information and the second time stamp information, used for determining a first time delay (i.e. a delay time of the UWB base station from receiving the request frame to sending the response frame after processing). The fourth difference value is a difference value between the fifth time stamp information and the fourth time stamp information, used for determining a second time delay (i.e. a delay time of the UWB tag from receiving the response frame to sending the termination frame after processing).

[0042] In a specific implementation, the first time stamp information is , the fourth time stamp information is , the first difference value is represented as:

[0043] ;

[0044] ;

[0045] wherein the first difference value is represented as ; the fourth time stamp information is represented as ; the first time stamp information is represented as ; the signal propagation time is represented as ; and the third difference value (the first time delay) is represented as .

[0046] The third time stamp information is , the sixth time stamp information is , the second difference value is represented as:

[0047] ;

[0048] ;

[0049] wherein the second difference value is represented as ; the sixth time stamp information is represented as ; the third time stamp information is represented as ; the signal propagation time is represented as ; and the fourth difference value (the second time delay) is represented as .

[0050] The second time stamp information is , the third difference value (the first time delay) is represented as:

[0051] ;

[0052] wherein the is a third difference value (first time delay); the is a third time stamp information; the is a second time stamp information.

[0053] The fifth time stamp information is , the fourth difference value (second time delay) is expressed as:

[0054] ;

[0055] wherein the is a fourth difference value (second time delay); the is a fifth time stamp information; the is a fourth time stamp information.

[0056] The signal propagation time is expressed as:

[0057] ;

[0058] wherein the is a signal propagation time; the is a first difference value; the is a second difference value; the is a third difference value (first time delay); the is a fourth difference value (second time delay).

[0059] The distance of the UWB tag to the UWB base station is expressed as:

[0060] ;

[0061] wherein the is the distance of the UWB tag to the UWB base station; the is a signal propagation time; the is the propagation speed of electromagnetic wave in air.

[0062] The method provided by the embodiment eliminates the interference of the local clock offset of the UWB tag and the UWB base station and the signal processing delay (first delay and second delay) on the distance measurement, and only retains the actual flight time of the pulse signal in space. Specifically, in the first difference (the total time of the UWB tag from sending a request frame to receiving a response frame) and the second difference (the total time of the UWB base station from sending a response frame to receiving a termination frame), both the propagation time of the pulse signal back and forth between the UWB tag and the UWB base station and the delay of signal processing of both parties are mixed. By calculating the product of the first difference and the second difference, the product of the first delay and the second delay, and taking the difference between the two, the term related to the signal propagation time can be effectively separated. Dividing by the sum of the first difference, the second difference, the first delay and the second delay can further offset the error caused by the clock offset, and finally the accurate one-way signal propagation time is obtained. The advantage of this method is that it does not need to strictly synchronize the clocks of the tag and the base station, and can eliminate the influence of non-propagation delay on distance measurement, significantly improve the accuracy of distance measurement, provide a reliable foundation for subsequent indoor positioning based on distance information, and is especially suitable for scenes with high positioning accuracy requirements in complex indoor environments.

[0063] S103, iteratively optimizing the distance based on a Kalman filtering algorithm to obtain optimized distance information.

[0064] Specifically, the Kalman filtering algorithm is a recursive estimation algorithm based on probability statistics. By combining the system's dynamic model prediction and actual measurement data, the estimation of the system state is continuously corrected, and the state estimation result closer to the true value can be continuously output in the presence of noise and uncertainty. Its core process includes two iterative steps of prediction (predicting the current state based on the last state estimation and system model) and update (correcting the prediction result with the current measurement value).

[0065] It should be noted that the present application considers that the pulse signal is easily affected by environmental interference (such as additional delay caused by signal reflection and refraction due to multipath effect, signal attenuation or propagation path deviation caused by obstacles), measurement error of hardware devices (such as small deviation of timestamp record, delay fluctuation of signal transceiver circuit), etc. The directly calculated distance may contain noise or jump, resulting in insufficient accuracy and stability of distance data; and the Kalman filtering can effectively suppress noise and smooth the fluctuation of distance data by using the algorithm characteristics, continuously approaching the true distance through iterative optimization, providing more reliable basic data for subsequent distance-based positioning, tracking and other applications, and improving the accuracy and robustness of the overall system.

[0066] In the implementation, the Kalman filtering algorithm is used to iteratively optimize the distance to obtain the optimized distance information, including: taking the distance optimized at the previous time as the state estimation value, predicting the predicted value at the current time through a state transition equation, the state transition equation is constructed based on the dynamic characteristics of signal propagation, and the state estimation error covariance is updated at each iteration; the distance is taken as the state estimation value at the first iteration; the distance is taken as the measured value, and the Kalman gain is determined based on the ratio of the prediction error covariance to the observation noise covariance; the predicted value and the measured value are weighted and summed based on the Kalman gain to obtain the optimized distance value; the optimized distance value is equal to the predicted value plus the product of the Kalman gain and the difference between the measured value and the predicted value; the current optimized distance value is taken as the state estimation value at the next time, and the optimization process is repeatedly performed until the optimized distance value converges.

[0067] Specifically, the Kalman filter is initialized: the process noise covariance Q and the observation noise covariance R are set, and the state estimation error covariance P is initialized as a preset value. At the first iteration, the distance between the UWB base station and the UWB tag calculated in the previous step is taken as the state estimation value x. Further, according to the state transition equation, the predicted value x_pred at the current time is directly calculated using the last state estimation value, and the prediction error covariance is updated after each iteration:

[0068] p_pred = P + Q;

[0069] Wherein, the p_pred is the prediction error covariance at the current time; the P is the state estimation error covariance at the current time; the Q is the process noise covariance.

[0070] Further, the Kalman gain is calculated according to the ratio of the prediction error covariance p_pred and the observation noise covariance p_pred + R:

[0071] K = p_pred / (p_pred + R);

[0072] Wherein, the K is the Kalman gain; the p_pred is the prediction error covariance; the p_pred + R is the observation noise covariance.

[0073] The predicted value and the current measured value are combined to obtain the optimized distance value through weighted summation:

[0074] x = x_pred + K * (d - x_pred);

[0075] Wherein, the x is the optimized distance value; the x_pred is the predicted value; the K is the Kalman gain; the d is the measured value.

[0076] After each iteration, the state estimation error covariance is updated:

[0077] P = (1 - K) * p_pred;

[0078] wherein the P is the updated state estimation error covariance; the K is the Kalman gain; and the p_pred is the prediction error covariance. The current optimized distance value is taken as the state estimation value at the next time, and the prediction and updating steps are repeatedly executed until the optimized distance value converges or a preset iteration number is reached, to obtain the optimized distance information. It should be noted that the specific implementation process and implementation principle of the Kalman filtering algorithm can be referred to the description in the related art, which will not be described here.

[0079] The method provided in this embodiment considers that, in terms of signal propagation, the UWB pulse signal is susceptible to the complex indoor environment (such as multipath effect, obstacle shielding), resulting in random fluctuations and systematic deviations in the directly calculated distance, while the Kalman filter dynamically models the signal propagation through the state transition equation, can predict the distance change under the normal signal propagation path, and filter abnormal fluctuations; in terms of noise processing, the introduction of the observation noise covariance quantifies the hardware measurement error (such as timestamp recording deviation), the prediction error covariance reflects the uncertainty of the prediction model, and the ratio of the two dynamically adjusts the Kalman gain, so that the filtering process can quickly respond to the real distance change (such as tag movement) and effectively suppress noise interference (such as false distance caused by signal reflection); in terms of iterative optimization, each iteration takes the optimization result of the last round as a new starting point, forms a closed-loop feedback mechanism, and as the number of iterations increases, the error covariance gradually decreases, indicating that the confidence of the system in state estimation is continuously improved, until it converges to a stable value, and the finally output optimized distance information not only retains the real-time of the measured value, but also filters noise through dynamic prediction and adaptive weighting, improves the stability and accuracy of the distance data, and is more suitable for subsequent scenarios that require high-precision distance information.

[0080] In S104, a positioning model calculates the position information of the UWB tag based on the optimized distance information and the position coordinates of the UWB base station, and sends the position information to the smart phone of the vehicle owner.

[0081] Specifically, the positioning model comprises an input module, a distance matching module, a connection module, a graph convolution network module and an output module. The input module, as a data inlet, classifies and receives and pre-processes the optimized distance information and the position coordinates of the UWB base station, providing basic data for subsequent calculation. The distance matching module introduces the importance matching weight of the distance between the UWB tag and the UWB base station, and adjusts the optimized distance information by weighting, highlights the distance relationship between the key base station and the tag through weight distribution, suppresses the influence of noise or abnormal data, and improves the reliability of the distance data, providing more accurate input for subsequent positioning calculation. The connection module splices the adjusted distance value output by the distance matching module with the base station coordinates provided by the base station position input module in the dimension, generates a connection vector value containing distance and position double information, and provides a structured input for the graph convolution network. The graph convolution network module constructs an adjacency matrix based on the continuous positioning points on the owner's action trajectory, further generates a node embedding matrix (mapping spatial relationship to a low-dimensional vector space), and mines the implicit geometric relationship between the UWB base station and the UWB tag through graph convolution operation, finally outputs the accurate position information of the UWB tag, and realizes the mapping conversion from multi-dimensional features to three-dimensional coordinates.

[0082] In a specific implementation, the positioning model calculates the position information of the UWB tag based on the optimized distance information and the position coordinates of the UWB base station, comprising:

[0083] (1) The input module receives the optimized distance information and the position coordinates of the UWB base station; the input module comprises a base station position input module and a distance input module, the base station position input module receives the position coordinates of the UWB base station, and the distance input module receives the optimized distance information.

[0084] Specifically, the input module comprises two parts: a base station position input module and a distance input module. The base station position input module receives the position coordinates of the UWB base station, and the distance input module receives the distance information optimized by the Kalman filtering algorithm. Since the position of each base station is based on a two-dimensional coordinate system, including the horizontal coordinate and the vertical coordinate, the base station position input module has 2N items in total, and the distance input module has N items in total, N being the number of UWB base stations.

[0085] (2) The distance matching module performs weight matching on the distance based on the importance of the distance between the UWB tag and the UWB base station, adjusts the optimized distance information based on the weight, and obtains the adjusted distance value.

[0086] In a specific implementation, the distance matching module performs weight matching on the distances based on the importance of the distances between the UWB tags and the UWB base stations, adjusts the optimized distance information based on the weights, and obtains adjusted distance values, including: determining the importance of the distances between the UWB tags and the UWB base stations based on the difference between the optimized distance information and the real distances; the smaller the difference, the higher the importance; mapping the importance to parameter values to construct a weight matrix; the higher the importance, the larger the parameter values corresponding to the combination, and the higher the corresponding weight; performing linear transformation on the optimized distance information by using the trained weight matrix, performing weighted summation on the distance information according to the importance of different distance information, and obtaining adjusted distance values.

[0087] Specifically, first, the number of UWB tags and UWB base stations is set, and the weight matrix is initialized. Further, the original ranging data is input into the graph convolutional neural network, feature propagation and transformation are performed through the graph convolutional layer, the difference (loss) between the distance information predicted by the model (optimized distance) and the real distance is calculated, the gradient of the weight matrix is calculated based on the loss function (such as MSE) through the back propagation algorithm. The optimizer (such as Adam) automatically updates the parameters of the weight matrix according to the gradient. The process of updating the weight matrix can be referred to the description in the related art, which will not be described here. The trained weight matrix and the optimized distance information are multiplied element by element, and the distance of all base stations corresponding to each UWB tag is weighted and summed to obtain the adjusted distance value. The adjusted distance value is represented as:

[0088] ;

[0089] Among them, the adjusted distance value is represented as: the weight matrix is represented as: and the optimized distance information is represented as: .

[0090] In a possible implementation, the weight matrix can be replaced by a Hadamard code vector, and the specific training process and implementation principle of the Hadamard code vector are similar to those of the weight matrix, which will not be described here. The adjusted distance value is represented as:

[0091] ;

[0092] Among them, the adjusted distance value is represented as: the Hadamard code vector is represented as: and the optimized distance information is represented as: .

[0093] The method provided in this embodiment adjusts the optimized distance information based on the importance of the distance between UWB tags and UWB base stations, which can significantly improve positioning accuracy and robustness. By learning the importance of the distance between UWB tags and UWB base stations, the contribution of reliable measurement pairs (tag-base station combinations with small differences) can be identified and amplified, while unreliable measurements affected by multipath effects or non-line-of-sight can be suppressed. By mapping importance to a weight matrix and applying a linear transformation, the influence of each distance measurement value can be dynamically adjusted, making the final positioning result more biased towards high-quality data. This adaptive weighting mechanism retains the ability of Kalman filtering to suppress temporal noise and introduces spatial dimension measurement quality assessment, forming a spatiotemporal dual optimization architecture. Specifically, in complex indoor environments, there may be obstacles or strong reflections between the tag and some base stations, causing the directly measured distance to deviate significantly from the true value. The weight adjustment based on importance can automatically reduce the weight of such abnormal measurements, avoiding them from misleading the positioning results. When the tag is close to a specific base station, the measurement accuracy of that base station is usually higher (stronger signal strength, simpler propagation path), and the weight adjustment mechanism will correspondingly increase its contribution to the final positioning, enhancing local positioning accuracy. Furthermore, the training process of the weight matrix can be iteratively optimized by incorporating historical data to further adapt to dynamically changing environmental characteristics (such as signal obstruction caused by population movement).

[0094] (3) The connection module splices the adjusted distance value with the location coordinates of the UWB base station to obtain the connection vector value.

[0095] Specifically, the link vector value refers to the feature vector generated by concatenating the adjusted distance information with the location coordinates of the UWB base stations in terms of dimensions. It is used to integrate the spatial geometric relationship between the two heterogeneous data. For each UWB tag, its adjusted distance values ​​to N base stations (vectors of length N) are concatenated sequentially with the two-dimensional coordinates of each base station (vectors of length 2N) to form a link vector of length 3N.

[0096] In practice, for each UWB tag-UWB base station pair, the adjusted distance information of the pair is concatenated with the location coordinates of the corresponding UWB base station along the dimensional direction to generate an intermediate feature matrix. Further, the intermediate feature matrix is ​​flattened into a three-dimensional vector to obtain the connection vector values.

[0097] The link vector value is represented as:

[0098] ;

[0099] Among them, the For connecting vector values; the The location coordinates of the UWB base station; This is the adjusted distance value; This is a connection function.

[0100] (4) The graph convolution network module constructs an adjacency matrix based on the continuous positioning points on the owner action trajectory, constructs a node embedding matrix based on the connection vector values and the adjacency matrix, and determines the position information of the UWB tag based on the node embedding matrix.

[0101] The graph convolution network module comprises an input submodule, a graph convolution submodule and an output submodule connected in sequence.

[0102] Specifically, the input submodule serves as a data inlet and receives a preset number of continuous connection vector values in time series, arranges and constructs a feature matrix in time sequence, and constructs an adjacency matrix based on the connection vector values of adjacent time points, thereby converting time series data into graph structure data and providing a basis for subsequent graph convolution operations. The graph convolution submodule extracts the spatial feature relationship between nodes (connection vectors at each time point) through graph convolution operation based on the feature matrix and the adjacency matrix provided by the input submodule, generates a node embedding matrix, and then performs linear transformation on the embedding matrix through a fully connected layer to convert the graph structure features into intermediate feature variables, completing the abstraction and conversion of features. The output submodule performs nonlinear mapping on the intermediate feature variables to convert the abstract features into position coordinate information of the UWB tag, realizes the final mapping from time-space features to physical positions, and outputs the positioning result of the tag.

[0103] In a specific implementation, the graph convolution network module constructs an adjacency matrix based on the continuous positioning points on the owner action trajectory, constructs a node embedding matrix based on the connection vector values and the adjacency matrix, and determines the position information of the UWB tag based on the node embedding matrix, which comprises: the input submodule receives a preset number of continuous connection vector values in time series, constructs a feature matrix in time sequence, sets the connection vector values of adjacent time points in the feature matrix to 1 to obtain an adjacency matrix; the rows of the feature matrix are time series, and the columns of the feature matrix are connection vector values; the graph convolution submodule constructs a node embedding matrix based on the feature matrix and the adjacency matrix, performs linear transformation on the node embedding matrix through a fully connected layer to obtain intermediate feature variables; and the output submodule maps the intermediate feature variables to generate the position information of the UWB tag.

[0104] Specifically, it receives a preset number of consecutive link vector values ​​from the time series, each link vector value being a fixed-dimensional array. All link vector values ​​are stacked vertically in chronological order to form a feature matrix of dimension [time_steps, feature_dim]. Here, time_steps is the preset number, feature_dim is the dimension of the link vector values, and each element of the feature matrix corresponds to a link vector value at a given time point. Further, it creates an all-zero matrix of dimension [time_steps, time_steps], where the rows and columns of the matrix correspond to time points in the time series. Iterates through each time point in the all-zero matrix, setting the matrix elements corresponding to adjacent time points to 1; that is, setting the matrix positions (i, i+1) and (i+1, i) corresponding to time t and t+1 to 1, forming a bidirectional adjacency matrix.

[0105] For example, in one embodiment, by selecting The feature matrix obtained from consecutive location points in a time series is a A matrix of dimension . Therefore, the characteristic matrix Represented as:

[0106] ;

[0107] Among them, the The characteristic matrix; For the first The first time point Location coordinates of the UWB base station; For the first The first time point The distance between a UWB base station and a UWB tag.

[0108] The adjacency matrix is ​​represented as:

[0109] ;

[0110] Among them, the These are elements in the adjacency matrix.

[0111] Optionally, the graph convolution submodule constructs a node embedding matrix based on the feature matrix and the adjacency matrix, comprising: based on the unit matrix and the degree matrix of the graph, combining the power operation of the degree matrix with the adjacency matrix to obtain a normalized adjacency matrix; the degree matrix is a diagonal matrix consistent with the dimension of the adjacency matrix, the corresponding positions of the main diagonal line of the degree matrix are the number of elements with a value of 1 in the corresponding row or column of each connection vector in the adjacency matrix, and the non-diagonal line positions are all 0; multiplying the normalized adjacency matrix with the feature matrix and the first weight matrix, and performing activation function processing on the product result; multiplying the result processed by the activation function again with the normalized adjacency matrix and the second weight matrix, and processing the product result again using the activation function to obtain the node embedding matrix.

[0112] Specifically, in the graph neural network, the degree matrix of the graph is used to describe the diagonal matrix of the connection strength of each node in the graph, which is uniquely determined by the adjacency matrix. For a given adjacency matrix, the sum of the elements of each row of the adjacency matrix is calculated, and the sum of each row is taken as the diagonal element, and the remaining positions are filled with 0 to obtain the degree matrix of the graph. The graph convolution submodule includes two sequentially connected graph convolution layers and a fully connected layer.

[0113] In specific implementation, two trainable weight matrices are defined, which are the first weight matrix and the second weight matrix, and the initial values of the weights of the weight matrices are set. The sum of each row of the adjacency matrix (i.e. the number of connections of each node) is calculated, the sum of each row is taken as the main diagonal element of the diagonal matrix, and the non-main diagonal elements are 0, to construct the degree matrix of the graph. Further, the adjacency matrix is standardized: a self-loop is added to the adjacency matrix (adjacency matrix + unit matrix), the inverse square root of the degree matrix is calculated, and the standardization is realized through matrix multiplication to obtain the normalized adjacency matrix. The normalized adjacency matrix, the feature matrix and the first weight matrix are multiplied in the first graph convolution layer in turn, and the product result is applied to the ReLU activation function. The normalized adjacency matrix and the output of the first graph convolution layer and the second weight matrix are multiplied in the second graph convolution layer in turn, and the product result is applied to the ReLU activation function to obtain the node embedding matrix. Further, a fully connected layer is defined, the input dimension is specified as the last dimension of the node embedding matrix, and the output dimension (the dimension of the intermediate feature variable) is specified. The node embedding matrix is received, for example, the dimension of the node embedding matrix is [batch_size, num_nodes, input_dim]. The batch and node dimensions are merged into one dimension, linearly transformed through the fully connected layer, the original batch and node dimensions are restored, and the intermediate feature variable is obtained.

[0114] For example, in an embodiment, the node embedding matrix is represented as:

[0115] ;

[0116] wherein the is a node embedding matrix; is a standardized result, the is an identity matrix; the is an adjacency matrix; the is a degree matrix of the graph; the , are respectively a first weight matrix, a second weight matrix; the is a feature matrix; is an activation function.

[0117] The intermediate feature variable is represented as:

[0118] ;

[0119] wherein the is an intermediate feature variable; the is a node embedding matrix; the is a weight of a fully connected layer; the is a bias term of the fully connected layer.

[0120] The method provided by the embodiment realizes effective modeling of the graph structure data through multiple means: the introduction of the identity matrix enhances the self-loop information, ensures that each node can retain its own features, the use of the degree matrix performs standardization processing, balances the influence of different connection degree nodes, and avoids the excessive dominance of the center node on information propagation, through two graph convolution operations (in combination with different weight matrices and activation functions), a multi-layer nonlinear transformation of the features is realized, the expression ability of the model on complex relationships is enhanced, and the repeated multiplication operation of the adjacency matrix and the feature matrix enables the nodes to aggregate neighbor information and capture local and global structural features of the graph. These means jointly act, so that the model can not only retain the physical meaning (such as distance and position information) of the original features in effect, but also can mine the hidden spatial correlation through the graph structure, improve the positioning accuracy and robustness, and is especially suitable for processing dynamic changes of the UWB network topology and multipath interference problems in a complex environment.

[0121] Further, the output submodule receives the dimension of the intermediate feature variable, sets the output dimension (default 3, corresponding to three-dimensional coordinates x, y, and z). A multi-layer fully connected network is constructed as a mapping layer, and the mapping layer structure is: the first layer: linear transformation, mapping the dimension of the intermediate feature variable to 64 dimensions, and the activation function: ReLU, introducing a nonlinear property. The second layer: linear transformation, mapping 64 dimensions to 32 dimensions, and the activation function: ReLU. The output layer: linear transformation, mapping 32 dimensions to the output dimension. After inputting the intermediate feature variable, the intermediate feature variable is transformed through each layer in turn, and finally the position information of the UWB tag is output.

[0122] The method provided by the embodiment significantly improves the precision and robustness of UWB tag positioning through multi-stage processing of the graph convolution network module: the input sub-module constructs the concatenated vector in time sequence into an adjacency matrix, converts the UWB positioning problem into a graph structure modeling, and captures the time continuity of the tag movement; the graph convolution sub-module balances the information propagation weight at different time points by standardizing the adjacency matrix, and combines two nonlinear transformations with a fully connected layer, which not only preserves the physical meaning of the original distance and position features, but also deeply mines the implicit spatial correlation in the time sequence, and enhances the robustness to multipath effect and non-line-of-sight interference; the output sub-module maps the abstract features to three-dimensional coordinates, realizing accurate conversion from complex signal features to physical positions. This architecture realizes spatio-temporal joint optimization in effect: in the time dimension, the dependence between adjacent time points is modeled by the adjacency matrix, smoothing the positioning jumps caused by signal fluctuations; in the spatial dimension, the graph convolution aggregates the information of multiple base stations, solving the uncertainty of single base station measurement.

[0123] The method provided by the embodiment, in the first aspect, from the whole process, high-precision positioning is realized through the collaborative design of "distance measurement-optimization-neural network processing": first, the UWB tag sends a pulse signal to no less than three base stations with known positions, and preliminarily measures the distance based on the time of flight, providing basic data for positioning; then, the Kalman filter smoothes the distance measurement noise through iterative optimization (prediction-update mechanism), reduces the interference such as multipath effect, and makes the distance information closer to the true value; finally, the positioning model integrates the optimized distance and base station coordinates, and through the weight adjustment of the distance matching module, the spatio-temporal features are spliced by the connection module, and the time sequence and spatial correlation are mined by the graph convolution network, and finally the accurate position is output. This whole process design forms spatio-temporal double optimization, which not only retains the real-time of the measured data, but also suppresses noise and mines implicit relationships through algorithms, and overall improves the positioning precision and environmental adaptability.

[0124] In the second aspect, high-precision distance measurement is realized through multi-frame interaction and timestamp calculation: the UWB tag and the base station transmit the first to the sixth timestamps through the request frame, the response frame and the termination frame, separate the real signal propagation time by calculating the time difference and eliminating the first time delay and the second time delay, and then obtain the distance combined with the speed of electromagnetic wave. This method can eliminate hardware delay and clock offset interference without strict clock synchronization, solves the problem of large error in traditional ToF distance measurement, and provides reliable original distance data for subsequent positioning.

[0125] In a third aspect, the positioning model improves positioning robustness through multi-layer feature processing: the input module classifies received data to ensure integrity; the distance matching module constructs a weight matrix based on distance importance, dynamically adjusts distance values, and highlights the contribution of reliable base stations; the connection module splices the adjusted distance and base station coordinates to integrate spatial and temporal features; the graph convolution network constructs a graph structure from the time-series connection vector, mines time continuity and spatial correlation through a standardized adjacency matrix and multi-layer convolution, and finally maps to location information. This processing both suppresses abnormal data through weight adjustment and captures spatio-temporal implicit relationships through graph convolution, significantly improving positioning accuracy and stability in complex environments.

[0126] Corresponding to the foregoing embodiment of the indoor positioning method for a car owner based on UWB technology, the present application also provides an embodiment of an indoor positioning system for a car owner based on UWB technology.

[0127] Figure 3 A structural schematic diagram of the indoor positioning system for a car owner based on UWB technology provided in Embodiment Two of the present application is shown in FIG. 2. Referring to FIG. 2, Figure 3 The indoor positioning system for a car owner based on UWB technology provided in the present embodiment includes a positioning server, a gateway, a plurality of UWB base stations, and a UWB tag installed in a car owner's smartphone;

[0128] The car owner's smartphone is configured to determine a UWB tag when there is a positioning demand, and send a positioning command to the UWB tag. Each car owner's smartphone corresponds to one UWB tag, and the UWB tag is the target to be positioned.

[0129] The UWB tag is configured to respond to the car owner's positioning demand, send a pulse signal to a plurality of UWB base stations with known positions after receiving the positioning command, and determine the distance from the UWB tag to the UWB base stations based on the time of flight of the pulse signal. The number of UWB base stations is not less than 3.

[0130] The positioning server is configured to iteratively optimize the distances based on a Kalman filter algorithm to obtain optimized distance information.

[0131] The positioning server is further configured to calculate the position information of the UWB tag based on the optimized distance information and the position coordinates of the UWB base stations.

[0132] The gateway is configured to receive distance information feedback from the UWB base stations and forward it to the positioning server, and also configured to send the position information calculated by the positioning server to the car owner's smartphone.

[0133] Optionally, the gateway transmits data to the UWB base station in wireless or wired form; the gateway transmits data to the UWB tag in wireless form; the gateway communicates data with the positioning server in wired form.

[0134] Specifically, the UWB-based indoor positioning system for the car owner comprises a positioning server, a gateway, N UWB base stations fixedly arranged in the room and having known positions, and a UWB tag installed in the car owner's mobile phone. The positioning server comprises a database, a memory, a processor, and a computer program stored in the memory and executable on the processor. The database is located in the memory and is used to store the position information of the UWB base stations and the UWB tag and the distance information between them. The computer program is used to implement the UWB-based indoor positioning method for the car owner. The gateway is a switch or a router, which is used to receive data packets from the UWB base stations or the UWB tag and forward the data packets to the positioning server. The gateway transmits data to the UWB base station in wireless or wired form; the gateway transmits data to the UWB tag in wireless form; the gateway communicates data with the positioning server in wired form.

[0135] The system of the embodiment can be used to execute the steps of the method embodiment, and the specific implementation principle and implementation process are similar, which will not be described here. Figure 1 The steps of the method embodiment are shown, and the specific implementation principle and implementation process are similar, which will not be described here.

[0136] The above is only the preferred embodiment of the present application, and is not used to limit the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the protection scope of the present application.

Claims

1. A method for indoor positioning of a vehicle owner based on UWB technology, characterized in that, The method comprises: When there is a positioning requirement, determining a UWB tag based on a car owner's smartphone, and sending a positioning command to the UWB tag through the smartphone; each car owner's smartphone corresponds to one UWB tag, and the UWB tag is a target to be positioned; After the UWB tag receives the positioning command, sending a pulse signal to a plurality of UWB base stations with known positions, and determining the distance from the UWB tag to the UWB base stations based on the time of flight of the pulse signal; the number of UWB base stations is not less than 3; Iteratively optimizing the distance based on a Kalman filtering algorithm to obtain optimized distance information; A positioning model calculates the position information of the UWB tag based on the optimized distance information and the position coordinates of the UWB base stations, and sends the position information to the car owner's smartphone.

2. The method of claim 1, wherein, The positioning model comprises an input module, a distance matching module, a connection module, a graph convolution network module and an output module; the positioning model calculates the position information of the UWB tag based on the optimized distance information and the position coordinates of the UWB base stations, comprising: The input module receives the optimized distance information and the position coordinates of the UWB base stations; the input module comprises a base station position input module and a distance input module, the base station position input module receives the position coordinates of the UWB base stations, and the distance input module receives the optimized distance information; The distance matching module performs weight matching on the distance based on the importance of the distance between the UWB tag and the UWB base station, adjusts the optimized distance information based on the weight, and obtains an adjusted distance value; The connection module splices the adjusted distance value and the position coordinates of the UWB base stations to obtain a connection vector value; The graph convolution network module constructs an adjacency matrix based on consecutive positioning points on the car owner's action trajectory, constructs a node embedding matrix based on the connection vector value and the adjacency matrix, and determines the position information of the UWB tag based on the node embedding matrix.

3. The method of claim 2, wherein, The distance matching module performs weight matching on the distance based on the importance of the distance between the UWB tag and the UWB base station, adjusts the optimized distance information based on the weight, and obtains an adjusted distance value, comprising: Based on the difference between the optimized distance information and the true distance, the importance of the distance between the UWB tag and the UWB base station is determined; the smaller the difference, the higher the importance; Mapping the importance to parameter values to construct a weight matrix; the higher the importance, the larger the parameter value corresponding to the combination, and the higher the weight corresponding to the combination; Using the trained weight matrix to perform linear transformation on the optimized distance information, and performing weighted summation on the distance information according to the importance of different distance information to obtain an adjusted distance value.

4. The method of claim 2, wherein, The graph convolution network module comprises an input submodule, a graph convolution submodule and an output submodule connected in sequence; the graph convolution network module constructs an adjacency matrix based on continuous positioning points on the owner action trajectory, constructs a node embedding matrix based on the link vector value and the adjacency matrix, and determines the position information of the UWB tag based on the node embedding matrix, comprising: The input submodule receives a preset number of continuous link vector values in time sequence, constructs a feature matrix in time sequence, sets the link vector values of adjacent time points in the feature matrix to 1 to obtain an adjacency matrix; the rows of the feature matrix are time sequences, and the columns of the feature matrix are link vector values; The graph convolution submodule constructs a node embedding matrix based on the feature matrix and the adjacency matrix, and linearly changes the node embedding matrix through a fully connected layer to obtain an intermediate feature variable; The output submodule maps the intermediate feature variable to generate the position information of the UWB tag.

5. The method of claim 4, wherein, The graph convolution submodule constructs a node embedding matrix based on the feature matrix and the adjacency matrix, comprising: Based on the unit matrix and the degree matrix of the graph, the power operation of the degree matrix is combined with the adjacency matrix to obtain a standardized adjacency matrix; the degree matrix is a diagonal matrix consistent with the dimension of the adjacency matrix, and the main diagonal line of the degree matrix corresponds to the number of elements with a value of 1 in the corresponding row or column of each link vector value in the adjacency matrix, and the non-diagonal line positions are all 0; The standardized adjacency matrix is multiplied by the feature matrix and the first weight matrix, and the product result is processed by an activation function; The result processed by the activation function is multiplied by the standardized adjacency matrix and the second weight matrix again, and the product result is processed by an activation function again to obtain a node embedding matrix.

6. The method of claim 1, wherein, After receiving the positioning command, the UWB tag sends a pulse signal to a plurality of UWB base stations with known positions, determines the distance from the UWB tag to the UWB base station based on the time of flight of the pulse signal, comprising: The UWB tag puts the first timestamp information and tag information when the positioning command is received into a request frame, and sends the request frame to a plurality of UWB base stations with known positions; After receiving the request frame, the UWB base station records the second timestamp information when the request frame is received, puts the third timestamp information after the first time delay and the base station information into a response frame, and sends the response frame back to the UWB tag; After receiving the response frame, the UWB tag records the fourth timestamp information when the response frame is received, puts the fifth timestamp information after the second time delay and the tag information into a termination frame, and sends the termination frame to the UWB base station; After receiving the termination frame, the UWB base station records the sixth timestamp information when the termination frame is received; Each UWB base station calculates the signal propagation time between each UWB base station and the UWB tag according to each timestamp information and the tag information, and determines the distance from the UWB tag to the UWB base station based on the signal propagation time.

7. The method of claim 6, wherein, Each of the UWB base stations calculates the signal propagation time between each UWB base station and the UWB tag according to the respective timestamp information and the tag information, determines the distance from the UWB tag to the UWB base station based on the signal propagation time, and comprises: calculating a first difference value of the fourth timestamp information and the first timestamp information, the first difference value being the time difference between the UWB tag sending a request frame and receiving a response frame; calculating a second difference value of the sixth timestamp information and the third timestamp information, the second difference value being the time difference between the UWB base station sending a response frame and receiving a termination frame; determining a first time delay based on a third difference value of the third timestamp information and the second timestamp information; determining a second time delay based on a fourth difference value of the fifth timestamp information and the fourth timestamp information; calculating a first product of the first difference value and the second difference value, calculating a second product of the first time delay and the second time delay, and calculating a fifth difference value of the first product and the second product; calculating the sum of the first difference value, the second difference value, the first time delay and the second time delay, and determining the quotient of the fifth difference value and the sum as the signal propagation time; the signal propagation time is the time for a pulse signal to reach the UWB tag from the UWB base station; determining the distance from the UWB tag to the UWB base station based on the product of the signal propagation time and the propagation speed of electromagnetic waves in air.

8. The method of claim 1, wherein, The distance is iteratively optimized based on the Kalman filtering algorithm to obtain optimized distance information, comprising: taking the optimized distance at the previous moment as the state estimation value, predicting the predicted value at the current moment through a state transition equation, and updating the state estimation error covariance at each iteration; the state transition equation is constructed based on the dynamic characteristics of signal propagation, and the distance is taken as the state estimation value at the first iteration; taking the distance as the measured value, and determining the Kalman gain based on the ratio of the prediction error covariance to the observation noise covariance; performing weighted summation on the predicted value and the measured value based on the Kalman gain to obtain the optimized distance value; the optimized distance value is equal to the predicted value plus the product of the Kalman gain and the difference between the measured value and the predicted value; taking the current optimized distance value as the state estimation value at the next moment, and repeating the optimization process until the optimized distance value converges.

9. A car owner indoor positioning system based on UWB technology, characterized in that, The car owner indoor positioning system based on the UWB technology comprises a positioning server, a gateway, a plurality of UWB base stations, and a UWB tag installed in a car owner's smartphone; wherein the car owner's smartphone is configured to determine the UWB tag when there is a positioning demand, and to send a positioning command to the UWB tag; each car owner's smartphone corresponds to one UWB tag, and the UWB tag is the target to be positioned; the UWB tag is configured to respond to the car owner's positioning demand, to send a pulse signal to a plurality of UWB base stations with known positions after receiving the positioning command, and to determine the distance from the UWB tag to the UWB base station based on the time of flight of the pulse signal; the number of UWB base stations is not less than 3. The positioning server is configured to perform iterative optimization on the distance based on a Kalman filtering algorithm to obtain optimized distance information. The positioning server is further configured to calculate position information of the UWB tag based on the optimized distance information and position coordinates of the UWB base station. The gateway is configured to receive distance information fed back by the UWB base station and forward the distance information to the positioning server, and further configured to send the position information calculated by the positioning server to the smart phone of the vehicle owner. 10.The UWB-based indoor positioning system for a vehicle owner according to claim 9, wherein, Data transmission between the gateway and the UWB base station can be in a wireless form or a wired form; data transmission between the gateway and the UWB tag can only be in a wireless form; and data communication between the gateway and the positioning server is in a wired form.

Images